THE COACHMAN (pulls the coach door open): Wachen Sie auf! Horen Sie! (two lights in the darkness draw attention to themselves.) Herr, wachen Sie doch auf, es is Zeit! (Shouting.) Es ist die hochste Zeit! (From the coach door a gentleman in a dark coat emerges, as if still half asleep.) Vaino Vahing, Madis Koiv. Faehlmann 1. Towards analytic poetics Analytic philosophy (i.e. philosophy that uses logic as its method) that was created by Gottlob Frege in the linguistic turn embraces the idea that language research or semiotics can be divided (according to the well-known distinction by Charles Morris) into three branches--syntax, semantics and pragmatics. The first of them is concerned with the interrelationships of expressions themselves, the second covers their relationships with the world and meanings, the third, the relations apparent in the usage of expressions and meanings. Naturally, this is but an abstract differentiation: although an analysis of the usage of literary language should first and foremost be the province of a pragmatics of poetic language usage (should such a discipline emerge more distinctly some day), no poetics can be conceived of without the data of syntax or semantics. It is within the limits of these two that most of the research has been carried out, with elements belonging to pragmatics having been drawn along intuitively. On the other hand, speech act theory does not belong to the narrow field of linguistic pragmatics, as meaning and usage are bound up inseparably and logically--rather, what we have at hand is general semantics in the framework of semiotics (see also Vanderveken 1990: 65-75). Poetic speech is more complicated and higher than ordinary speech, a language usage governed by numerous new restrictions that is often referred to (e.g. Lotman 1972: 18-23) as a secondary modelling system above or underlying the primary one. Jakobsonian treatments (see Jakobson 1960) clearly show the simultaneous superimposed position, as well as logical precedence of the poetic order as a particular kind of axiomatics as to everyday language usage. Of course, it is against the background of ordinary language as the primary system that literary language appears as secondarily modelling; an even more general semiotic approach would rather prefer a hierarchy of three stages. When we unite the natural and conventional meanings ([meaning.sub.NN]) differentiated by Grice (Grice 1989c) and the notions of the tertiary modelling system developed by Thomas Sebeok (Sebeok 1994) we would get the following scheme of functioning of a sign system. [FIGURE 1 OMITTED] An analysis of the poetic language usage shows that the theory of speech acts should and can be greatly refined to describe speech of the rhetorical type. But it is possible to do it well, for the illocutionary logic, being a well-formed theory by nature, allows for a successful addition of new formal layers--everything should be definable, reducible and transparent, at least in principle. Study of literature that departs from these premises, i.e. linguistic criticism (q.v. Fowler 1986) would without difficulties fulfil the requirement of scientific quality in the sense of Immanuel Kant. 2. The theory of speech acts I have been trying to introduce the speech act theory into the Estonian-language mindspace, setting the developing of a respective conceptual network for poetics as my further aim (see Merilai 1998, 2001). As is known, the speech act theory was launched by John L. Austin (see Austin 1984 (1982)), yet it was in the philosophy of John R. Searle (see Searle 1992 (1969)) that it gained its established standard formulation. In co-operation with Searle, Daniel Vanderweken began to develop illocutive logic (see Searle, Vanderveken 1985, Vanderveken 1990, 1991). The theory divides simple speech acts into five categories: assertives, i.e. saying how things are (A), comissives, i. …